Recognizing and analyzing variability in amyloid formation kinetics: Simulation and statistical methods.

We examine the phenomenon of variability in the kinetics of amyloid formation and detail methods for its simulation, identification and analysis. Simulated data, reflecting intrinsic variability, were produced using rate constants, randomly sampled from a pre-defined distribution, as parameters in an irreversible nucleation-growth kinetic model. Simulated kinetic traces were reduced in complexity through description in terms of three characteristic parameters. Practical methods for assessing convergence of the reduced parameter distributions were introduced and a bootstrap procedure was applied to determine convergence for different levels of intrinsic variation. Statistical methods for assessing the significance of shifts in parameter distributions, relating to either change in parameter mean or distribution shape, were tested. Robust methods for analyzing and interpreting kinetic data possessing significant intrinsic variance will allow greater scrutiny of the effects of anti-amyloid compounds in drug trials. Crown